Longitudinal study of adherence to anticoagulation guidelines in patients with atrial fibrillation

Abstract

Abstract Introduction Anticoagulation (AC) reduces the risk of stroke and systemic embolism (SSE) in patients with Atrial Fibrillation (AF). However, the association between changes in AF prevalence, risk factors for SSE, uptake of AC and incidence of SSE has not been documented specifically in these patients within a national population. Purpose Identifying antithrombotic prescribing and evaluating associations between changes in AF prevalence, SSE risk factors (CHA2DS2-VASC score) and uptake of AC. Evaluating relationships between changes in proportion of AF patients treated with AC and SSE rate at a population level. Developing a modelling tool which estimates rates of SSE (past and future) based on these factors, which can be used to drive improvements in health care. Method AF patients were identified in the population of Wales (SAIL databank) between 2012–2018. Temporal trends of AF, CHA2DS2-VASC scores, anti-thrombotic prescriptions and SSE events were evaluated. Multi-state Markov models were used to estimate SSE rates adjusted for AC and CHA2DS2-VASC scores. Simulation methods modelled SSE outcomes for the subsequent 7-years based on differing proportional population AC coverage. Results AF prevalence increased from 51,492 to 64,852 from 2012–18, with mean CHA2DS2-VASC score increasing from 3.0 to 3.9. AC prescription coverage increased (24,892 [48.3%] to 44,195 [68.1%]), whilst antiplatelet therapy alone or no antithrombotic therapy decreased (14,532 [28.0%] to 5,385 [8.3%] and 26,602 [52.0%] to 21,164 [33.0%] respectively). Hospitalisation rate for SSE in AF population decreased by over 20%, from 1,039 per 100,000 patients/quarter in 2012 to 809 per 100,000 patients/quarter in 2018. Markov modelling demonstrated a 39% lower SSE rate with AC compared to no AC over time, after adjustment for individual CHA2DS2-VASC (HR=0.61, 95% CI [0.58, 0.63]). Using the estimated progression rates, simulation models shown that an expected 3,574 SSE events per 100,000 per year could have been reduced to 2,956 (17% reduction rate) if AC adherence had been at the now recommended 90% rate from 2012. This model also predicts that improving AC coverage to 90% of patients over the next 7-years would reduce SSE rates by 12% per year (assuming no further increase in mean CHA2DS2-VASC score). Conclusion Despite the increase in both prevalence of AF and CHA2DS2-VASC between 2012–18, we observed a progressive decrease in SSE events along with increasing AC prescribing. Our study suggests not only that improved AC coverage is associated with better clinical outcomes, but also that the rate of therapeutic implementation is likely a crucial factor. Reducing transition time between evidence-based guideline publication and widespread clinical uptake of the recommendations appears to be an important opportunity to improve clinical outcomes at a population level and should inform healthcare policy development and implementation. Funding Acknowledgement Type of funding source: Other. Main funding source(s): Funded by research grant